edge detection in digital image using morphology operation
DESCRIPTION
This paper shows a method in digital image processing technique to find the defects in tablets. In this paper we use mathematical manipulation, to detect the defected tablet packet.TRANSCRIPT
ICRTEDC-2014 58
Vol. 1, Spl. Issue 2 (May, 2014) e-ISSN: 1694-2310 | p-ISSN: 1694-2426GV/ICRTEDC/14
EDGE DETECTION IN DIGITAL IMAGEUSING MORPHOLOGY OPERATION
Virdeep KaurAssistant Professor, St. Soldier Group of Institutions, Punjab, India
Abstract: Medicines have helped to make our lives easy.Drug industry is developing industry in terms ofproduction as well as consumption. Medication hasbecome very important in everyone’s life as we areaffected by so many diseases. But these medicines mightbe defected, tablets may be broken, there may be missingtablet in a strip and consumption of these drugs might bedangerous. This paper shows a method in digital imageprocessing technique to find the defects in tablets. In thispaper we use mathematical manipulation, to detect thedefected tablet packet.
Keywords: Digital Image Processing, Morphologyopening, Preprocessing.
I. INTRODUCTIONDigital image processing techniques and algorithms areapplied on images in order to remove error. In this paper,we use digital image processing technique to detect thebroken tablet. Such tablets are harmful to consume andmay have many side effects. The inspection process iseffective to detect the defects in tablets. Mathematicalmanipulation is used to detect the defect. This is done inMatlab10. First Image is taken and is converted into grayand then to binary and then noise is removed. Morphologyoperation is used to remove the noise. Morphologyoperation is applied on binary images therefore for thisimage is first converted into gray. This technique will findthe defect in those tablets which are circular in shape.
Figure 1 Different tablet packet with defect
II. METHODOLOGYIn this paper,statistical method is used that ismathematically values are calculated.Tablet strip consistsof various shapes like circular, rectangular, ellipse. In thispaper,we have taken circular ones. We know that thatcircular tablets have there particular area so if tablets arebroken they deviate from the roundness. So in this waydefect can be detected.
III. PROPOSED WORKIn this paper,statistical method is used to find the defect intablets.In order to find the defect,rgb image is convertedinto gray and then to binary. The binary image have noiseso in order to remove noise, morphology opening is used.Boundaries of the output is detected after pre processing.After this process determine the roundness of tablet,findarea and perimeter of each tablet. After calculating areaand perimeter,find the metric.Metric closer to 1 indicatesthat tablet is not broken or is completely round.area = πr
Where ‘r’ is the radius of a circular tabletperimeter=π×d
Where‘d’is the diameter of tabletTherefore, metric=4*pi*area/perimeter^2So,metric closer to 1 indicates that tablet is not defected.Also no. of tablets are calculated which helps to determinethe correct figure in a tablet strip. Counting can also help tofind the pharmaceutical company. They can simply discardthe strip containing less or greater no. of tablets.
Steps:1. Capture argb image.2. Convert the image into grayand then to binary.3. Remove the noise using morphology opening.4. Make the boundaries of all tablets.5. Label the matrix.6. Calculate the no. of tablets in a strip.7. Now, compute the roundness of tablets.8. Display the result,matrix equal to 1 means the tablet is
round and is without any defect.However,matrix not equalto 1 means tablet is defected.
59 ICRTEDC -2014
Input Image Gray Image
Image with Noise Cleaned Image
Metrics closer to 1 indicate that the tablet is approximately round
No. of Tablets Found:10
0.95
0.96
0.95
0.94 0.95
0.890.95
0.95
0.94
0.95
0
1000
2000
3000
4000
5000
6000
0 50 100 150 200 250
Figure 3 block diagram shows how to find the defect
IV. RESULTS
Figure 2
Defected Tablet Packet
Figure 3 Histogram of tablet packet
Figure 4 Removal of noise
Figure 5 Edge Detection
ICRTEDC-2014 60
Input Image Gray Image
Metrics closer to 1 indicate that the tablet is approximately round
No. of Tablets Found:9
0.96
0.95
0.95
0.96
0.94 0.95
0.95
0.95
0.94
Figure 6 Screenshot of time elapsed
Input rgb image which is having defected tablet isfirst converted into gray image as shown in figure 2.Histogram of the tablet packet is shown infigure3.gray image is then converted into binaryimage but the image is having some noise.Noise isremoved by applying morphology opening.Figure 4 shows the cleaned image after removal ofnoise.Figure 5 shows the edges of tablets.Figure 6shows the tablet whose metric is less than 1 is notround and is defected.
Figure 7 Tablet packet with missing tablet
Figure 8 one missing tablet
This method is also applied on tablet packet with missingtablet Figure 8 shows the tablet packet with one missingtablet. The tablet strip is of 10 tablet and the result showsonly 9 tablets .This indicates that one tablet is missing in
tablet packet.The whole process takes 2 Seconds oftime.Figure 9 shows the screen shot after the wholeprogramme was executed.LinuxDebian 5 has greaterspeed,so it is recommended to use this machine foroperation.
V. CONCLUSIONThis paper presents the application in drug industry usingdifferent techniques in digital image processing. It isimplemented in matlab10 software. This paper mentionsstatistical method for detecting the anomalies in drugs.Inthis paper,defect of tablet is detected using area of atablet,the tablets which deviatefrom the said area isconsidered as defected. We have taken the circular tablet,however this method can also be implemented on differentshapes of tablet packet using their area. Insteadmorphology operation used for noise removal, any othermethod can be used.For efficient work,the time takenshould be minimum.
REFERENCES[1] Abha Sharma1,SugandhaArora “ Inspection and
Classification of Defects in Pharmaceutical Capsules UsingNeural Network” International Journal of EngineeringResearch and Development ISSN: 2278-067X, Volume 1,Issue 10 (June 2012), PP.41-45
[2] ŞtefanOprea, IoanLiţă, Mariana Jurianu, DanieAlexandruVişan, Ion BogdanCioc, “Digital Image ProcessingApplied in Drugs Industry for Detection of Broken AspirinTablets” Electronics, Communications and ComputersDepartment, University of Pitesti Str.Targul din Vale Nr.Pitesti, Romania,2008
[3] F Anthony C.karleff, Neil E.Scott& Robert Muscedure,Flexible Design for a cost effective,high throughput Inspectionsystem for pharmaceutical capsules. IEEE, 2008
[4] Dong-Su Kim, Wang-Heon Lee , In-So Kweon, “Automaticedge detection using 3x3 ideal binary pixel patterns and fuzzy-based edge thresholding,” in Pattern Recognition Letters in2004
[5] El-Khamy,S.E., Lotfy,M., El-Yamany,N.A.,“A ModifiedFuzzy Sobel Edge Detector”,17. National Radio ScienceConference, Feb.22-24 2000, Minufiya University, Egypt
[6] Miosso, C.J., Bauchspiess, A., “Fuzzy Inference SystemApplied to Edge Detection in Digital Images”, Proceeding ofthe Brazilian Conference on Neural Networks, pp. 481-486,2001
[7] HardeepKaur and Er.NidhiGarg, “Inspection of DefectivePharmaceutical Capsules using Harris Algorithm”,International Journal of Advances in ElectronicsEngineering,2009
[8] R.Thilepa “A Paper On Automatic Fabrics Fault ProcessingUsing Image Processing Technique in Matlab”, Signal &Image Processing : An International Journal(SIPIJ) Vol.1,No.2, December 2010.
[9] Peng Zhao and Shutao Li “Tablets Vision InspectionApproach Using Fourier Descriptors and Support VectorMachines” The 9th International Conference for YoungComputer Scientists,2008.
[9] Ramya.S,,Suchitra.J,Nadesh R.K”Detection of BrokenPharmaceutical Drugs using Enhanced Feature ExtractionTechnique”
[10] Nain, Neeta, et al. "Morphological Edge Detection andCorner Detection Algorithm Using Chain Encoding." IPCV 6(2006): 520-525.
[11] TizhooshH.R.,“Fast fuzzy edge detection”, Proceedings ofFuzzy Information Processing Society, 2002, pp. 239-242.
[12] Abdallah A. Alshennawy, And Ayman A. Aly, "EdgeDetection In Digital Images Using Fuzzy Logic Technique", Inproceedings of World Academy of Science Engineering andTechnology, Vol.51 , pp.178-186, 2009.
61 ICRTEDC -2014
[13] Kaur, ErKiranpreet, ErVikramMutenja, and ErInderjeetSingh Gill. "Fuzzy Logic Based Image Edge DetectionAlgorithm in MATLAB." International Journal of ComputerApplications 1.22 (2010): 55-58.
[14] Baig, Hasan,Jeong-A. Lee, and Jieun Lee. "Performanceevaluation of CPU-GPU and CPU-only algorithms fordetecting defective tablets through morphological imagingtechniques." Information Systems andTechnologies(CISTI),2012 7th Iberian Conference on.IEEE,2012.